An experimental approach to Mathematics

Today I read an interesting article that I am delighted to share with you. These thoughts are written by a prominent (Argentine-American) computer scientist - Gregory Chaitin. It is published by Inference Review:

• G. Chaitin. Doing Mathematics Differently, Inference Review, 2(1), 2016

It is devoted to Author’s baby - the Algorithmic Information Theory (AIT). The issues discussed there should be of interest for any working mathematician. Perhaps, this is my favourite passage from the aforementioned article:

What if we take Gödel incompleteness very seriously and throw away rigor? Suppose you have a property of the prime numbers which has been checked on the computer. You graph it and there is a beautiful curve, and it is fit beautifully by a very simple equation. What if you cannot prove it? A physicist would publish anyway. But a pure mathematician does not care how much empirical evidence there is, or how accurately this simple formula fits the curve. You need a proof!